The mediating role of SARS-CoV-2 variants between income and hospitalization due to COVID-19: a period-based mediation analysis

The mechanisms facilitating the relationship between low income and COVID-19 severity have not been partitioned in the presence of SARS-CoV-2 variants of concern (VOCs). To address this, we used causal mediation analysis to quantify the possible mediating role infection with VOC has on the relationship between neighborhood income (exposure) and hospitalization due to COVID-19 among cases (outcome). A population-based cohort of 65 629 individuals residing in British Columbia, Canada, was divided into 3 periods of VOC co-circulation in the 2021 calendar year, whereby each period included co-circulation of an emerging and an established VOC. Each cohort was subjected to g-formula mediation techniques to decompose the relationship between exposure and outcome into total, direct, and indirect effects. In the mediation analysis, the total effects indicated that low income was associated with increased odds of hospitalization across all periods. Further decomposition of the effects revealed that income is directly and indirectly associated with hospitalization. The resulting indirect effect through VOC accounted for approximately between 6% and 13% of the total effect of income on hospitalization. This study underscores, conditional on the analysis, the importance of addressing underlying inequities to mitigate the disproportionate impact on historically marginalized communities by adopting an equity lens as central to pandemic preparedness and response from the onset.